The swf should hit breakpoints in the sort code. Is it not doing that?
I can usually use fdb on the swf.
Sent from my LG G3, an AT&T 4G LTE smartphone
------ Original message------
From: Mihai Chira
Date: Sat, Jul 18, 2015 12:49 PM
To: dev@flex.apache.org;
Subject:Re: git commit: [flex-sdk] [refs/heads/develop] - FLEX-34854 FLEX-34884
Adding unit test to make sure that the changes in FLEX-34854 haven't changed
anything about findItem.
Now I'm stuck because although I can run the swf that was generated at
the previous step in debug mode in IntelliJ, the most further
breakpoint which I can catch is in the TestCase.as constructor. Which
implies, to me, that the test case in the mxml file is defined and
created, but not executed. Do I need to change the code to get it to
execute somehow (maybe using that static function init()?), and then
regenerate the SWF?
On 18 July 2015 at 21:39, Mihai Chira <mihai.ch...@gmail.com> wrote:
> Yep, it's pretty clear which ones are still failing.
>
> I'm looking, for instance, at
> SF_module_sparkButtonBar_global_tester.mxml, whose only test case
> should run perfectly. I made a sample app which does (seemingly) the
> same thing, and the Sort and SortField do return the correct locale
> value ("ru-RU"). So in order to figure out why in mustella it's
> failing I'd need to debug it. I'm following the instructions on the
> wiki now.
>
> FYI while doing so I got an error about cygwinshell:
> "java.io.IOException: Cannot run program "${cygwinshell}":
> CreateProcess error=2, The system cannot find the file specified". And
> when I trace it with <echo> I get
>
> [echo] shell = ${cygwinshell}
> [echo] cygwinshell = ${cygwinshell}
>
> It worked only when I changed [1] into [2] in mustella/build.xml
> (basically to use the git shell).
>
>
> [1] <available file="/cygwin/bin/sh.exe" property="cygwinshell"
> value="/cygwin/bin/sh.exe"/>
> <available file="/cygwin64/bin/sh.exe" property="cygwinshell"
> value="/cygwin64/bin/sh.exe"/>
>
> [2] <property name="cygwinshell" value="c:\Program Files
> (x86)\Git\bin\sh.exe"/>
>
>
> On 17 July 2015 at 15:09, Alex Harui <aha...@adobe.com> wrote:
>> I wasn’t able to reset mustella until just now so it might be hard to find
>> your results in the 100’s of errors in build # 1533. Hopefully build #
>> 1534 will be cleaner.
>>
>> FWIW, my magic incantation that resets mustella is to log into the server,
>> open ApacheFlex/dependencies/FlashPlayer_Debug and click on and close each
>> debug player. No idea why it works. Maybe Tom can try that some day to
>> see if it works for him as well.
>>
>> -Alex
>>
>> On 7/17/15, 3:16 AM, "Mihai Chira" <mihai.ch...@gmail.com> wrote:
>>
>>>Cool, thanks for the link. I'll take a look tonight or tomorrow at the
>>>results.
>>>On 16 Jul 2015 22:53, "Alex Harui" <aha...@adobe.com> wrote:
>>>
>>>> The link at [1] is the build server that is currently down. It doesn’t
>>>> run mustella.
>>>>
>>>> Mustella tests are at http://flex-mustella.cloudapp.net and failures are
>>>> sent to commits@
>>>>
>>>> Every couple of days, we start getting weird failures that requires
>>>> fiddling with the mustella server. It looks like we’re in that funky
>>>> state right now so I will try to deal with that in about 3 hours so the
>>>> next run after that should show the effects of your changes.
>>>>
>>>> -Alex
>>>>
>>>> On 7/16/15, 1:48 PM, "Mihai Chira" <mihai.ch...@gmail.com> wrote:
>>>>
>>>> >At least one of the failures should now be fixed. If others are left,
>>>> >I'll tackle them in the next few days.
>>>> >How can I monitor the test results? The bookmark I have for the build
>>>> >server is this[1] and it doesn't seem to be working. Sorry if there
>>>> >was a url change that I didn't catch.
>>>> >
>>>> >[1] http://apacheflexbuild.cloudapp.net:8080/job/flex-sdk_test/
>>>> >
>>>> >On 16 July 2015 at 21:52, Mihai Chira <mihai.ch...@gmail.com> wrote:
>>>> >>> changes. Can you take a look?
>>>> >>
>>>> >> on it
>>>>
>>>>
>>
